权益分类	功能权益	普通用户	{{item.name}}会员
{{category.name}}	{{benefitItem.name}}

Active Interaction Considering Spatiotemporal Arrangement of Incormation

考虑信息时空安排的主动交互

基本信息

批准号：
12650369
负责人：
KENAKO Masahide
金额：
$ 1.28万
依托单位：
The University of Electro-Communications
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (C)
财政年份：
2000
资助国家：
日本
起止时间：
2000 至 2001
项目状态：
已结题

项目摘要

(1) Intelligent representation of imagesSeveral intelligent representation methods of visual information have been devised according to content of information, user' s requirement, type of application and so on. For example, a method to increase resolution of viewed area when a user gazes it more than predetermined time, and another method to increase resolution along the movement of view direction.(2) Active interaction considering spatiotemporal arrangement of informationA new type of natural interaction between a user and a computer has been studied. A target object is indicated by words representing spatial arrangement and temporal order. Various words indicating spatiotemporal arrangements are listed up. Since most of these words contain ambiguity, spatiotemporal range represented by each word is estimated by Fuzzy theory and history of interaction, Experimental system followed this idea has proved that a user can pick up a specific object in a computer screen as if one would give instructions to other people in a real world. Furthermore by learning the individual characteristics, it is possible to choose a target object more efficiently.(3) Anthropomorphic agent using facial caricatureAnthropomorphic agent using facial caricature of a specific person whom a user knows very well will improve the human-computer interaction much better. A new synthesis method of facial caricature has been devised, which utilizes eigenspace of each of various facial features. In order to realize an animated facial caricature, two important methods have been studied. One is to control the change of mouth shape synchronized with speech. The other is to represent facial expressions.

(1)图像的智能表示根据信息内容、用户需求、应用类型等，设计了几种视觉信息的智能表示方法，例如，当用户注视观察区域超过预定时间时提高观察区域分辨率的方法，以及沿观察方向沿着移动提高分辨率的方法。(2)考虑信息时空排列的主动交互研究了一种新型的用户与计算机之间的自然交互。目标对象由表示空间布置和时间顺序的词指示。列出了各种表示时空安排的词语。由于这些词大多含有歧义，每个词所代表的时空范围是由模糊理论和历史的互动估计，实验系统遵循这一思想已经证明，用户可以拿起一个特定的对象在计算机屏幕上，如果一个人会给其他人在一个真实的世界的指令。此外，通过学习个体特征，可以更有效地选择目标对象。(3)使用面部漫画的拟人代理使用用户非常熟悉的特定人的面部漫画的拟人代理将更好地改善人机交互。提出了一种新的人脸漫画合成方法，该方法充分利用了各种人脸特征的特征空间。为了实现动画人脸漫画，研究了两种重要的方法。一是控制口型变化与说话同步。另一种是表现面部表情。